Unraveling the characteristics of early esophageal neuroendocrine carcinoma using multi-model endoscopy: A retrospective study of serial cases

Abstract

BACKGROUND

Early esophageal neuroendocrine carcinoma (E-NEC) is a rare but aggressive malignancy with poorly understood endoscopic features. Despite advancements in multi-model endoscopy, including white light endoscopy, magnifying endoscopy narrow-band imaging (NBI), and endoscopic ultrasonography (EUS), the diagnostic characteristics of early E-NEC remain unclear. Comprehensive evaluation using these techniques can improve early detection and guide clinical management. This study aimed to investigate the endoscopic features of early E-NEC using multiple imaging modalities. We hypothesized that specific endoscopic patterns, such as irregular microvascular morphology or signs of submucosal invasion, could reliably distinguish early E-NEC from other esophageal lesions.

AIM

To characterize early E-NEC using multi-model endoscopy and identify diagnostic endoscopic features.

METHODS

Clinical data of four patients with esophageal submucosal lesions identified by gastroscopy and pathologically diagnosed as E-NEC in the Department of Gastroenterology, The First Affiliated Hospital of Zhejiang University School of Medicine between January 2020 and August 2024 were retrospectively analyzed, and their manifestations under multi-model endoscopy were observed. Grayscale values of ultrasound images in three patients with E-NEC and eight with esophageal leiomyoma were calculated using Image J software and compared using the Mann-Whitney U test.

RESULTS

Among the four patients with early E-NEC, two were males and two were females, with ages ranging from 60-70 years. White light endoscopy revealed three patients with mild local ulceration on the lesion surface. NBI images were available in two patients, revealing the intraepithelial papillary capillary loop of B1-B2 and B2 types. EUS was performed in three patients, all showing solitary hypoechoic lesions originating from the muscularis mucosa/mucosal muscular layer or submucosa; peripherally enlarged lymph node was observed in one patient. Average grayscale value of EUS images for early E-NEC (n = 3, 63.79 ± 1.42) was significantly higher than that for esophageal leiomyoma (n = 8, 49.44 ± 11.57; P = 0.01).

CONCLUSION

Endoscopic features of early E-NEC differ from those of esophageal leiomyoma, including NBI imaging, lymph node metastasis, and a notably higher echo intensity on EUS.

Key Words: Esophagus; Carcinoma; Neuroendocrine; Endoscopic ultrasonography; Leiomyoma; Narrow band imaging

Core Tip: This was a retrospective observational study that investigated the endoscopic characteristics of early esophageal neuroendocrine carcinoma (E-NEC) and how they differed from those of other esophageal submucosal tumors. This study analyzed four patients with early E-NEC using multi-model endoscopy and revealed key diagnostic differences from esophageal leiomyomas, including higher echogenicity on endoscopic ultrasonography, irregular microvascular patterns on narrow-band imaging, and lymph node enlargement. These findings highlight the potential of combined endoscopic modalities to improve the early detection and diagnostic accuracy of E-NEC.

INTRODUCTION

Neuroendocrine neoplasms (NENs) are a rare heterogeneous group of epithelial tumors that originate from peptidergic neurons and neuroendocrine cells, and are characterized by neuroendocrine differentiation. NENs include well-differentiated neuroendocrine tumors (NETs), poorly differentiated neuroendocrine carcinomas (NECs), and mixed neuroendocrine-non-NENs[1], which exhibit a wide range of biological behaviors, from indolent, slow-growing lesions to highly aggressive and metastatic malignancies[2]. NENs have the capacity for whole-body distribution, with gastroenteropancreatic NENs comprising approximately 60%-75% of cases, followed by those in the lungs and mediastinum[3]. Esophageal NENs (E-NENs) are extremely rare, accounting for only approximately 1.6% of all NENs and occurring far less frequently than other gastrointestinal NENs[4]. However, with advancements in early cancer screening and diagnostic techniques, the incidence of E-NENs has gradually increased in recent decades[5]. Among E-NENs, < 1% are well-differentiated NETs, with the vast majority being poorly differentiated small cell NECs[6,7]. Nevertheless, esophageal NEC (E-NEC) accounts for only 0.05%-7.6% of all esophageal malignant tumors[8,9], and is characterized by aggressive behavior, rapid progression, and early metastasis. E-NEC primarily affects middle-aged and older individuals, with a higher prevalence in males[10].

The tumor-node-metastasis staging system of E-NEC mirrors that of esophageal squamous cell carcinoma (ESCC)[11,12]. Early E-NEC is defined as lesions limited to the mucosal or submucosal layer (stages T1a and T1b), with or without nodal metastasis. Clinically, we found that early E-NEC often presents with absent or non-specific symptoms, such as mild dysphagia or retrosternal discomfort, which are easily overlooked or mistaken for benign esophageal conditions such as leiomyoma. Conventional imaging and tumor markers frequently fail to detect early lesions, and endoscopic biopsy results may be inconclusive due to submucosal tumor growth and the overlying normal epithelium[13]. As a result, E-NEC is frequently diagnosed at an advanced stage (31%-90% of cases) when regional lymph node or distant metastases have already occurred, contributing to its poor prognosis[2,13,14]. Studies have reported a median survival of approximately 11 months and 5-year survival rate of < 10%[14]. Given these challenges, early detection using endoscopy is essential for improving patient outcomes.

Conventional white light endoscopy (WLE) uses broad-spectrum white light to visualize the gastrointestinal mucosa, and is currently the standard modality for initial screening and surveillance because of its high-resolution, real-time imaging capabilities for anatomical structures[15]. However, WLE has limited sensitivity for detecting premalignant and early-stage flat malignant lesions such as intraepithelial neoplasia and early cancer. Early E-NECs, which typically present as subtle mucosal or submucosal lesions, may therefore be missed on standard WLE. These diagnostic limitations can be addressed by integrating advanced techniques such as narrow-band imaging (NBI) and endoscopic ultrasound (EUS). NBI utilizes filtered blue (415 nm) and green (540 nm) lights to enhance visualization of the microvascular architecture[16]. By exploiting the absorption characteristics of hemoglobin, NBI improves contrast, highlights abnormal vascular and mucosal patterns, aids in the differentiation of benign and malignant lesions, and enhances the detection of early neoplasia.

Early E-NEC typically appears as a hypoechoic lesion originating from the mucosa, submucosa, or muscularis propria[17]. Although EUS is recommended by the European Society of Gastrointestinal Endoscopy as an optimal tool for characterizing subepithelial lesions because of its ability to assess lesion location, size, layer of origin, echogenicity, morphology, and surrounding structures[18], it cannot reliably differentiate all subepithelial lesion types when used independently. Early E-NEC often appears similar to esophageal leiomyoma on EUS, leading to misdiagnosis, poor prognosis, and delayed treatment[2]. Therefore, this study aimed to investigate the characteristics of early E-NEC using multimodal endoscopic imaging techniques, including WLE, NBI, and EUS, and to compare these findings, especially the gray value of a lesion under EUS, with those of esophageal leiomyoma to enhance diagnostic accuracy and reduce diagnostic pitfalls. Additionally, this article reviews the literature on recent advances in the diagnosis, treatment, and pathogenesis of early E-NEC.

MATERIALS AND METHODS

Study design and patients

We retrospectively reviewed patients with submucosal esophageal lesions detected by gastroscopy at the Department of Gastroenterology, The First Affiliated Hospital of Zhejiang University School of Medicine between January 2020 and August 2024. Within the review period, four patients diagnosed with NEC using diagnostic endoscopic submucosal dissection (ESD) or surgical resection were included. Patients’ clinical profiles, laboratory and imaging findings, endoscopic observations, and pathological results were collected. This study was approved by the Ethics Committee of the First Affiliated Hospital of Zhejiang University School of Medicine.

Statistical analyses

Image analysis software Image J was used to calculate the gray values of the images of EUS. Clinical parameters of patients were expressed as mean ± SD of continuous variables and comparisons between groups were presented using the Mann-Whitney U test. Statistical analysis was performed using SPSS version 26.0 statistical software. Statistical significance was considered at P < 0.05.

Literature review

For review of the literature, database search was performed in PubMed, with the following search algorithm, [“neuroendocrine carcinoma” (All Fields) OR “NEC” (All Fields)] AND [“oesophagus” (All Fields) OR “esophagus” (All Fields) OR esophageal (All Fields)]. A literature search was conducted for articles published up to the year 2025. Studies focusing on early E-NEC were included, whereas reviews without full texts were excluded.

RESULTS

Clinical symptoms

Clinical characteristics of the four patients with early E-NEC are summarized in Table 1. Age of the patients ranged between 60-70 years (mean 66 years; median 67 years). Presenting symptoms included dysphagia in one patient, recurrent retrosternal discomfort in two patients, and no symptoms in one patient. Tumor marker levels were within normal limits in all the patients. Contrast-enhanced computed tomography (CT) showed mild wall thickening in the mid-esophagus in one patient, while no significant abnormalities were detected in the remaining three patients.

Table 1 Clinical data and endoscopic findings of 4 cases with early esophageal neuroendocrine carcinoma.

Case	Sex	Age (year)	Presentation	Examination	Location (cm)	Size (cm)
1	Male	60	No symptoms	Tumor markers were normal; contrast-enhanced CT showed no definite mass	30	0.94 × 0.36
2	Male	64	Retrosternal discomfort	Tumor markers were normal; contrast-enhanced CT showed no definite mass	27	1.3 × 0.9
3	Female	70	Recurrent retrosternal discomfort with abdominal distension and chest tightness	Tumor markers were normal; contrast-enhanced CT revealed mild mucosal enhancement in the mid-esophagus and a small nodule in the anterior mediastinum	28	1.2 × 0.7
4	Female	70	Dysphagia	Tumor markers were normal; contrast-enhanced CT showed no definite mass	30	0.5 × 0.6

CT: Computed tomography.

Endoscopic manifestations

Under WLE, three of the four patients with early E-NEC exhibited localized mild ulceration on the lesion surface, while one presented with a smooth-surfaced polypoid lesion (Figure 1). The lesions were located at 27 cm, 28 cm, 30 cm, and 30 cm from the incisors and involved the mid to lower esophagus. The lesion sizes were between 0.5 cm × 0.6 cm and 1.2 cm × 1.1 cm. NBI was performed in two patients (Figure 2), which revealed irregular or dilated abnormal microvessels. Based on the Japan Esophageal Society classification[19], the intraepithelial papillary capillary loop (IPCL) patterns were identified as types B1-B2 and type B2.

Figure 1 White light endoscopic images of cases 1 to 4. A: Case 1, a 1.0 cm mucosal elevation with central depression was observed 30 cm from the incisors. The lesion was mobile on palpation; B: Case 2, a 1.2 cm lesion was noted 27 cm from the incisors; C: Case 3, a type 2A mucosal elevation, measuring approximately 1.2 cm × 0.7 cm, was observed 28 cm from the incisors. The lesion exhibited surface erythema, pallor, irregular granularity, erosion, spontaneous bleeding, and well-defined margins; D: Case 4, a mucosal elevation measuring approximately 1.1 cm with slight central depression was detected 30 cm from the incisors.

Figure 2 Magnifying endoscopy with narrow band imaging of case 3 and case 4. A: Case 3, magnifying endoscopy with narrow band imaging revealed that the lesion at the area with redness was type B1-B2 vessels; B: Case 4, magnifying endoscopy with narrow band imaging revealed that the lesion at the area with redness was type B2 vessels.

Ultrasonographic endoscopic manifestations

EUS was routinely performed according to well-acknowledged procedures using an Olympus EU-ME2 (Olympus Corporation, Tokyo, Japan) and mini probe with a frequency of 12 MHz. Among the patients with early E-NEC, three underwent EUS (Figure 3), which revealed solitary hypoechoic lesions originating from the muscularis mucosae or submucosa with homogeneous internal echogenicity, clear borders, normal peripheral esophageal hierarchical architecture, and intact muscularis mucosae and epithelial membranes. In one patient, multiple enlarged lymph nodes were observed outside the esophageal wall. The average grayscale value of EUS images for early E-NEC (n = 3) was significantly higher than that of esophageal leiomyoma [n = 8, 63.79 ± 1.42 vs 49.44 ± 11.57, P = 0.01, Cohen’s d = 1.402; 95% confidence intervals (CI): -0.2718 to 3.077], indicating that early E-NEC lesions exhibited higher echogenicity than esophageal leiomyomas (Figure 4).

Figure 3 Endoscopic ultrasonography images of cases 1 to 3. A: Case 1, after water immersion, a 12 MHz miniprobe scan revealed a well demarcated, homogenous, hypoechoic lesion measuring 0.94 cm × 0.36 cm within the muscularis mucosae; B: Case 2, after water immersion, a 12 MHz miniprobe scan revealed a lesion primarily characterized by thickening of the second layer with a measured thickness of 0.27 cm. The layer structure remained distinct, with intact muscularis propria and tunica albuginea; C: Case 3, after water immersion, a 12 MHz miniprobe scan revealed a hypoechoic lesion primarily involving the mucosal layer and partially extending into the superficial submucosa, with hypoechoic structures noted outside the esophageal wall.

Figure 4 Endoscopic ultrasonography images of cases with esophageal leiomyoma. A: Case 1, a submucosal elevation was observed 25 cm from the incisors with a smooth surface. After water immersion, a 12 MHz miniprobe scan revealed a well-defined, homogenous, hypoechoic lesion originating from the muscularis mucosae, measuring approximately 0.39 cm × 0.60 cm in cross-section. The surrounding esophageal wall layers appeared normal; B: Case 2, a submucosal elevation was observed 27 cm from the incisors with a smooth surface. After water immersion, a 12 MHz miniprobe scan revealed a well-defined, homogenous, hypoechoic lesion originating from the muscularis mucosae, measuring approximately 0.65 cm × 0.40 cm in cross-section. The surrounding esophageal wall layers appeared normal; C: Case 3, conventional gastroscopy revealed a spindle-shaped submucosal elevation approximately 1.0 cm in length, located 33 cm from the incisors, with localized surface depression. After water immersion, a 12 MHz miniprobe scan showed a well-defined, homogenous, hypoechoic lesion originating from the muscularis mucosae, measuring approximately 0.59 cm × 0.36 cm in cross-section. The surrounding esophageal wall layers appeared normal.

DISCUSSION

E-NEC is an extremely rare malignancy characterized by rapid progression, high metastatic potential, and poor prognosis[20,21], and is influenced by tumor stage, grade, and distant metastases[6]. A multicenter retrospective study in Taiwan reported a median survival of 8.2 months (range: 0.2-104 months) among 38 patients with E-NEC[22]. A 10-year database-based retrospective cohort study further demonstrated that E-NEC carries a higher risk of metastasis [62.20% vs (58.72% and 46.66%)] and worse prognosis than ESCC and adenocarcinoma[9]. E-NEC predominantly affects individuals > 60 years of age, with a male-to-female ratio of (4-6):1, and is most commonly located in the middle and lower esophagus, particularly in the lower segment[9].

In this study, we found that early E-NEC was frequently misdiagnosed as esophageal leiomyoma in clinical practice. Esophageal leiomyoma, a common benign tumor of the esophagus, typically presents as a slow-growing, solitary, submucosal mass with expansive growth and predominantly occurs in the middle and lower esophagus. Endoscopy reveals a submucosal elevation with well-defined margins and similar color to the adjacent mucosa. EUS usually reveals a round or ovoid hypoechoic lesion, mostly originating from the muscularis propria of the esophagus, with a few arising from the muscularis mucosae[23].

Clinically, patients with early E-NEC are often asymptomatic or present with non-specific gastrointestinal symptoms similar to those of esophageal leiomyoma, such as dysphagia, retrosternal discomfort, etc.[10]. This lack of distinctive clinical features complicates the early diagnosis of E-NEC. Additionally, routine examinations, including tumor markers, typically present within normal limits, and contrast-enhanced CT shows no abnormalities, further mimicking the presentation of esophageal leiomyoma. Furthermore, E-NEC exhibits submucosal growth with normal epithelium, typically overlying the lesion[17], often resulting in negative or non-specific biopsy findings such as esophageal papillomatous hyperplasia[13]. These factors contribute to pathological diagnostic challenges and may delay accurate identification and timely disease management.

However, E-NEC is characterized by rapid progression and high metastatic potential. In this study, one patient initially presented with a 1.0 cm submucosal elevation at 30 cm from the incisors on conventional gastroscopy, with central depression, smooth mobility, and intact overlying mucosa. EUS suggested a submucosal lesion, likely a leiomyoma. Tumor markers were unremarkable, and contrast-enhanced CT showed no mass. Biopsy revealed squamous epithelial papillary hyperplasia, leading to a presumptive diagnosis of leiomyoma and follow-up recommendations. Six months later, repeat endoscopy revealed a large ulcerative lesion. Biopsy revealed small-cell E-NEC, and subsequent surgery confirmed lymph node metastasis. A retrospective review noted enlarged periesophageal lymph nodes on the initial EUS, which was overlooked. This case highlights the non-specific clinical features and auxiliary findings of early E-NECs and their endoscopic resemblance to leiomyomas. Diagnosis via biopsy can be challenging, and attention should be paid to extramural lymphadenopathy. These findings emphasize the importance of better characterization of endoscopic features of early E-NEC.

Therefore, three additional patients who were pathologically diagnosed with E-NEC at The First Affiliated Hospital of Zhejiang University School of Medicine were included in this study. Their clinical data and endoscopic findings were retrospectively analyzed. Notably, none of the patients were definitively diagnosed with E-NEC during initial gastroscopy and EUS, which once again proves the diagnostic challenges and certain limitations of the current EUS techniques in detecting E-NEC. For further investigation, we inductively analyzed the clinical and endoscopic characteristics of all the four patients with E-NEC and compared them with the EUS findings from the eight patients diagnosed as esophageal leiomyoma using ESD or endoscopic mucosal resection. The analysis revealed that early E-NEC exhibits distinct EUS echogenicity, is frequently associated with lymph node enlargement, and presents with specific vascular patterns under NBI. Therefore, we performed a grayscale analysis of EUS images using Image J software to compare the echogenicity between E-NECs and esophageal leiomyomas. The results revealed that E-NEC lesions exhibited significantly higher echogenicity, with a statistically significant difference in gray-scale analysis (P = 0.01). This finding may have important implications for gastrointestinal endoscopists to improve the diagnostic accuracy and differentiation of submucosal esophageal lesions.

Additionally, early E-NEC typically exhibits a dense and irregular microvascular network with twisted and dilated vessels, which appears under NBI as prominent brownish or tan vascular patterns in contrast to the adjacent normal tissue[24]. In contrast, esophageal leiomyomas generally show lower vascular density and more regular vascular structures, resulting in lighter or more regular microvascular images under NBI, with a less pronounced or minimal difference from the surrounding. In this study, NBI revealed type B1-B2 IPCL patterns in two patients, supporting the diagnostic utility of vascular features in differentiating E-NEC from benign lesions.

Furthermore, E-NEC is associated with a high rate of lymph node metastasis, whereas esophageal leiomyomas are benign and rarely metastasize[11,25]. Thus, enlarged periesophageal lymph nodes on EUS should raise a suspicion of E-NEC. Although this study identified several distinguishing features between E-NEC and esophageal leiomyoma, including differences in EUS echogenicity and NBI images, as well as lymph node metastasis, there are some limitations that should be considered. First, as E-NECs are extremely rare, we included only four patients in our study, which limited the ability to derive more significant statistical interpretations. Although the observed effect size (Cohen’s d = 1.402) indicated a large difference between the two groups, the 95%CI of the effect size included 0, which may raise concerns regarding the robustness of the effect. This finding should also be interpreted in the context of a small sample size, which is known to inflate uncertainty in estimating effect sizes and widen CI, especially in studies with unequal group sizes. Importantly, despite the wide interval, the intergroup difference remained statistically significant (P = 0.01), suggesting that the observed effect was unlikely to be due to random chance. We believe that the inclusion of 0 in the CI reflects limited statistical power rather than the absence of a true effect. Thus, future clinical studies with larger sample sizes are needed to validate these observations and more precisely determine the true magnitude of the observed effect.

Second, our study employed offline analysis using Image J software to compare echogenicity between E-NECs and esophageal leiomyomas. Although this method allows quantitative assessment of gray-scale values, it has several limitations that constrain its clinical applicability. First, because this post-processing approach requires exporting and manually processing images, it lacks real-time interaction with the operator, thereby limiting its utility in real-time diagnosis and intraoperative decision-making. Second, the lack of standardization in region of interest selection and filtering parameters, which often depend on operator experience, results in poor reproducibility and limits cross-center validation.

To address the aforementioned limitations, recent advances have extended artificial intelligence (AI)-integrated endoscopic systems beyond in vitro image analysis for classification to real-time detection of tissue characterization in early esophageal neoplasia[26]. Convolutional neural networks based on WLE and NBI have achieved high accuracy in detecting superficial ESCC, with a reported sensitivity and specificity of 84% and 73%, respectively[27]. Knabe et al[28] developed an EUS-AI system using a deep convolutional neural network for esophageal adenocarcinoma, demonstrating that AI could accurately identify mucosal (T1a) and submucosal (T1b) carcinomas, with high accuracy in differentiating between the two. This architecture can be adapted to differentiate early E-NECs from esophageal leiomyomas. Prospective clinical studies should account for the rarity of E-NECs. Few-shot learning may help overcome data scarcity. Furthermore, integrating multimodal inputs, such as WLE, NBI, and EUS lymph node metastasis with AI-based grayscale analysis, may enhance the accuracy of early E-NEC diagnosis.

For the literature review, 15 studies identified by the database search were eligible and included (Table 2)[13,24,29-41]. Multi-model endoscopy, which is typically used for the detection of early E-NEC, including WLE, NBI, and EUS, is summarized in Table 2. WLE remains the initial screening tool; however, its sensitivity for early flat and subepithelial lesions is limited. NBI has become increasingly valuable because it offers enhanced visualization of microvascular patterns and mucosal architecture. E-NECs often present with irregular IPCLs, which can be classified using the Japan Esophageal Society system to differentiate malignancies from benign lesions[19]. Additionally, EUS plays a critical role in assessing lesion depth and origin, particularly in submucosal tumors. Hypoechoic masses originating from the muscularis mucosa or submucosa with possible lymphadenopathy, raise the suspicion of E-NEC. Further quantitative analysis of EUS images, such as grayscale value comparison, may better distinguish E-NECs from benign subepithelial tumors such as leiomyomas.

Table 2 Review of all cases of early esophageal neuroendocrine carcinoma.

Case	Ref.	Age	Sex	Location	Presentation	NBI	EUS	Stage	Pathology	Treatment	Survival
1	[13]	77	M	Middle	No symptoms	Type B1, type R	The lesion was localized in the mucosa	T1aN0M0	NEC, SCC	ESD, surgery, CTx	No sign of disease recurrence 12 months after surgery
2	[24]	80	M	Middle/28 cm	No symptoms	Type B3, type R	A hypoechoic mass invading the third hyperechoic layer (submucosal layer)	cT1bN0M0/SM2	NEC, SCC	ESD, CTx, RTx	3 months lymph node metastasis; 6 months multiple liver metastases; 8 months died
3	[29]	51	M	25 cm	Dysphagia	NA	NA	T1aN0M0	NEC, squamous cell HGIN	ESD	No recurrence or metastasis 18 months
4	[30]	75	F	Middle	NA	Type B1	A hypoechoic lesion located in the mucosal layer and partially within the superficial submucosal layer	pT1aN0M0	NEC, SCC	ESD, CTx	60 months without any recurrence
5	[31]	49	F	35 cm	Intermittent epigastric soreness and heartburn	NA	A 44 mm × 3.3 mm, well demarcated, homogenous, hypoechoic, round mass lesion within the mucosal layer, and the submucosal layer beneath the lesion was observed to be intact	T1aN0M0	NEC	EMR	2 months alive
6	[32]	55	F	Upper	Intermittent mild dysphagia	NA	Showed that the bulged lesion was highly echoic and homogeneous, originating from the muscularis mucosa	SM2/T1aN0M0	NEC	ESD, CTx	Remained disease-free during a 2-year follow-up
7	[33]	65	M	27 cm	Retrosternal discomfort	NA	Characterized the lesion as an ovoid hypoechoic mass originating from the muscularis mucosae, displaying clear margins and homogeneous echogenicity	T2N1M0	NEC	ESD	Developed systemic metastases after 16 months
8	[34]	54	M	Lower	Dysphagia, dizziness, and melena, GERD	NA	NA	NA	NEC	CTx, surgery	20 months alive
9	[35]	62	M	Distal	Epigastric pain for 2 months; obvious difficulty in swallowing accompanied by chest pain 14 months later	NA	A low echo nodule in the muscularis mucosa, with un-uniform internal echo and a clear boundary. The nodule appeared to have a capsule and the submucosa was smooth and continuous	NA	NEC	CTx, RTx	Developed after 14 months, died of brain metastases 17 months later
	[35]	68	F	NA	Epigastric pain for 1 month	NA	A hypoechoic mass in the muscularis mucosa invading the submucosal layer, and the submucosal layer was thinner than normal but still continuous	T1bN0M0	NEC	ESD, surgery	9 months alive
10	[36]	79	M	Esophagogastric junction	No symptoms	NA	NA	T1a	NEC, EC	ESD, surgery	4 years alive
11	[37]	63	M	37 cm	No symptoms	NA	NA	cT1aSM1	NEC, SCC, EC	ESD	15 months alive
12	[38]	55	M	Middle	NA	Type V1, type V2, type V3	NA	SM2	NEC, SCC	ESD, CTx	55 months alive
13	[39]	61	M	Middle	Epigastralgia provoked by swallowing	NA	There was thickening of the second layer, but the SM layer was not infiltrated, and it was judged to be a lesion in the mucosa	SM1	NEC	ESD, CTx, surgery	Lymph node metastasis 14 months later; lung metastases 17 months later; died 25 months later
14	[40]	70	M	Lower/33 cm	Discomfort when swallowing, weight loss	Type 3, type 4IB	The lesion was depicted as a hypoechoic area in layer 2; the third layer was preserved and the depth of the lesion was considered to be intramucosal	pT1b (SM: 200 μm)	NEC, SCC	ESD, CTx	Paracardial lymph node swelling 5 months later; died 22 months later
15	[41]	63	M	33 cm	A loss of appetite	NA	A 7 mm hypoechoic nodule on the mucosal layer with slight invasion to the submucosal layer	T1bN0M0	NEC	EMR	18 months alive

M: Male; F: Female; NA: Not available; GERD: Gastro-esophageal reflux disease; NBI: Narrow band imaging; EUS: Endoscopic ultrasound; SM: Submucosa; T: Tumor; N: Node; M: Metastasis; NEC: Neuroendocrine carcinoma; SCC: Squamous cell carcinoma; HGIN: High-grade intraepithelial neoplasia; EC: Esophageal carcinoma; ESD: Endoscopic submucosal dissection; CTx: Chemotherapy; RTx: Radiotherapy; EMR: Endoscopic mucosal resection.

Histopathological confirmation of E-NEC remains essential, with biopsies showing small round cells with scant cytoplasm, high mitotic activity, and necrosis. Immunohistochemistry is indispensable for the definitive diagnosis of E-NEC, with positive markers, such as chromogranin A, synaptophysin, and cluster of differentiation 56, along with high Ki-67 proliferation indices, confirming the neuroendocrine nature and grading of a tumor[6]. However, owing to sampling limitations and the submucosal location of many E-NECs, biopsies may yield false negative results. Therefore, emerging technologies, such as confocal laser endomicroscopy, which allows real-time microscopic visualization of mucosal structures, and AI-assisted image analysis are under investigation for their potential to improve diagnostic precision[42-45]. Furthermore, the integration of circulating tumor biomarkers and molecular diagnostics, including next-generation sequencing, enables the identification of genomic alterations, such as tumor protein p53, retinoblastoma susceptibility gene (RB1), and notch receptor 1 mutations, offering potential diagnostic and therapeutic implications[46-50]. These advances collectively contribute to a more nuanced understanding of E-NEC and facilitate earlier and more accurate diagnosis. Nonetheless, challenges remain in distinguishing early E-NEC from mimicking lesions, such as poorly differentiated squamous cell carcinoma or leiomyoma, underscoring the need for continued refinement of diagnostic tools and the development of standardized endoscopic and molecular criteria.

Treatments and prognoses are listed in Table 2. Among the 16 patients reviewed, four underwent surgery, in which three received additional surgical treatment after postoperative pathological confirmation of E-NEC by ESD, and one patient underwent surgery after chemotherapy. All the patients achieved favorable outcomes. Nine patients underwent ESD without subsequent surgery, including three who underwent ESD alone, two who achieved good prognoses, and one who developed systemic metastases 16 months later. The remaining six patients received adjuvant chemotherapy or radiotherapy, resulting in mixed outcomes. Three patients had favorable responses, while three experienced disease progression and ultimately died. Specifically, one patient developed lymph node metastasis at 3 months, liver metastases at 6 months, and died at 8 months. Another patient underwent thoracotomy with esophagectomy and lymphadenectomy for lymph node metastasis at 14 months, developed lung and bone metastases at 17 months, and died at 25 months. The third patient developed paracardial lymph node swelling at 5 months and died at 22 months. Two patients underwent endoscopic mucosal resection with good outcomes. One patient who received chemoradiotherapy alone had a poor prognosis. No consensus on standardized treatment strategy for early E-NEC has been reached at present, although various treatment modalities, including endoscopic resection, surgery, chemotherapy, radiotherapy, biological therapy, and targeted therapy, have been described[51-54]. Traditionally, surgical resection has been the cornerstone of E-NEC treatment, particularly due to its aggressive nature and high propensity for early metastasis. Esophagectomy with regional lymphadenectomy is considered the standard approach for localized or resectable disease, providing comprehensive pathological assessment and regional control[11]. However, surgery is associated with significant morbidity, particularly in older adult patients and those with comorbidities, which limits its applicability in certain clinical contexts. In recent years, advances in endoscopic techniques have led to an increased interest in less invasive options such as ESD[29,33,35], especially for early E-NEC confined to the mucosa or superficial submucosa without lymphovascular invasion. ESD allows for en bloc resection with precise histological evaluation and organ preservation; however, it is limited to early lesions without nodal involvement. Therefore, individualized treatment plan is essential, with consideration of tumor stage, depth of invasion, histological grade, patient comorbidities, and preferences. A multidisciplinary approach is crucial for optimizing outcomes and appropriate integration of endoscopic, surgical, and systemic therapies[20,55].

The pathogenesis and molecular characteristics of E-NEC remain unclear, but are hypothesized to involve the transformation of pluripotent stem cells, neuroendocrine precursor cells, or malignant cells from various histologic subtypes in the esophageal epithelium[14,56,57]. Genomic studies have shown that genetic alterations commonly associated with NETs, such as tumor protein p53 and RB1 mutations, also play a role in E-NEC pathogenesis[58-60]. Mutations in notch receptor 1, FAT atypical cadherin 1, ARID3A, and nuclear factor erythroid 2-related factor 2 occur more frequently in E-NECs than in NECs from other gastrointestinal locations[61]. Significant somatic copy number variations have been investigated in E-ENC, including deletions at 13q14 (harboring RB1) and 3p12-14 (harboring fragile histidine triad and roundabout guidance receptor 1), and amplifications of cyclin E1 and MYC[60]. Additionally, epigenetic regulation of the p63 gene has been demonstrated to play an important role in the transdifferentiation of ESCC into NEC[62]. Commonly mutated signaling pathways in E-NEC include the Wnt-β-catenin, Notch, and ERBB pathways[61]. For example, low activity of the Notch signaling pathway was observed in E-NECs, which is characterized by downregulation of Notch receptors and upregulation of Notch antagonists[60].

From transcriptomic landscape, E-NEC shares similar gene expression with small cell lung cancer, but differs from ESCC and esophageal adenocarcinoma. The upregulated genes were involved in DNA replication, cell cycle, and neuroendocrine differentiation, whereas the downregulated genes were associated with cell adhesion. From an immune perspective, E-NECs exhibit a T cell-excluded phenotype, with insufficient T cell infiltration observed in most cases. A study revealed that cluster of differentiation 8 T cells failed to infiltrate a tumor parenchyma, instead, they aggregated in the surrounding stroma[60].

CONCLUSION

Early E-NEC and esophageal leiomyoma share similar clinical symptoms and endoscopic manifestations, often leading to missed or delayed diagnoses. However, the findings of the present study suggest that multimodal endoscopic imaging is a promising tool for differentiation. Suspicion of E-NEC should be increased in cases showing similar or slightly increased echogenicity on EUS, irregular microvascular patterns on NBI, or enlarged periesophageal lymph nodes. Such cases require close monitoring and proactive management. Further studies with larger cohorts are needed to validate these findings and enhance diagnostic accuracy in clinical practice.

ACKNOWLEDGEMENTS

We thank Farhin Shaheed Kalyani (School of Medicine, Zhejiang University) for kindly supporting this study.